The risk of bacterial infection of the endometrium causing uterine disease in cattle is increased in the progesterone-dominated luteal phase of the ovarian cycle, while oestrogens or oestrus are therapeutic or protective against disease. The first line of defence against bacteria, such as Escherichia coli that cause inflammation of the endometrium, is the innate immune system, which recognises bacterial lipopolysaccharide (LPS). This study tested the hypothesis that cyclic variation in ovarian hormone concentrations alters innate immune responses within the bovine endometrium. Ex vivo organ cultures of endometrium, and in vitro cultures of endometrial epithelial and stromal cells, and peripheral blood mononuclear cells (PBMCs), all mounted inflammatory responses to E. coli or LPS, with secretion of inflammatory mediators interleukin 1β (IL1β), IL6 and IL8, and increased expression of mRNA encoding IL1B, IL6, CXCL8 (IL8) and CCL5. However, these inflammatory responses, typical of innate immunity, were not affected by the stage of ovarian cycle in which the endometrium was collected for organ culture, or by exogenous oestradiol or progesterone. Although a dexamethasone-positive control reduced inflammation stimulated by E. coli or LPS, treatment with oestradiol or progesterone, or inhibitors of oestradiol or progesterone nuclear receptors, did not affect endometrial cell or PBMC secretion of IL1β, IL6 or IL8, or IL1B, IL6, CXCL8 and CCL5 gene expression. In conclusion, the stage of the oestrus cycle or ovarian steroids did not modulate the innate immune response in the bovine endometrium in vitro.
Microbial infections of the uterus are a common cause of infertility, abortion, pre-term labour and clinical disease of humans and animals (Wira et al. 2005, Turner et al. 2012). In dairy cows, post partum infection rates reach >90% after parturition, with clinical disease evident in nearly half of these cows. Disease of the uterus may persist for several weeks, is refractory to current treatments and leads to infertility (Sheldon et al. 2009). As a result, uterine disease of dairy cows is of economic importance, costing the EU dairy industry €1.4 billion/year (Sheldon et al. 2009). Escherichia coli are an important cause of pathology in the endometrium (Sheldon et al. 2010), with infection often preceding infection with Trueperella pyogenes and anaerobic bacteria (Sheldon et al. 2002). Furthermore, infection with E. coli is associated with negative effects on the ovary, hypothalamic–pituitary axis, and animal health and welfare (Williams et al. 2007). The endometrium forms an essential barrier to infection of the uterus. Cytokines and chemokines orchestrate the recruitment and activation of immune cells to combat invading pathogens (Wira et al. 2005). Responses to microbial infection depend on pattern recognition receptors, such as the Toll-like receptors (TLRs), which are expressed by cells of the endometrium (Herath et al. 2006, 2009a, Sheldon & Bromfield 2011). In particular, endometrial epithelial and stromal cell responses to E. coli infection are mediated by TLR4, which binds lipopolysaccharide (LPS) leading to secretion of the chemokine interleukin 8 (IL8) and the cytokine IL6 (Cronin et al. 2012).
The endometrium undergoes physiological changes under the control of the ovarian steroids oestradiol and progesterone to create an environment suitable for pregnancy (Lewis 2003, Wira et al. 2005). These steroids also have an impact on endometrial disease. During the follicular phase of the oestrus cycle, when oestradiol concentrations are high, the endometrium is more resistant to infection, while the progesterone-dominated luteal phase of the oestrus cycle is associated with a predisposition to development of disease (Rowson et al. 1953, Lewis 2003, 2004). Despite the clear effect of ovarian cycle on uterine disease progression, mechanistic data for the immune polarising effects of oestradiol and progesterone are less apparent. However, differences in uterine cellular profiles have been noted. Cells harvested from the uterine lumen around the time of ovulation secrete higher concentrations of cytokines and chemokines compared with cells harvested during the luteal phase of the oestrus cycle (Fischer et al. 2010). In some studies, ovarian steroids were associated with changes in neutrophil function (Roth et al. 1983), whereas in other studies, there were no consistent differences in peripheral leukocyte populations or their function (Subandrio & Noakes 1997, Subandrio et al. 2000, Winters et al. 2003). The inconsistency of leukocyte population differences and neutrophil functional changes suggest that steroid control of uterine disease progression may be the product of altered endometrial cell responses. Exogenous oestradiol and progesterone alter prostaglandin secretion in vivo in cows, sheep and pigs (Del Vecchio et al. 1992, Seals et al. 2002, Wulster-Radcliffe et al. 2003). In vitro, exogenous ovarian steroids reduce the secretion of prostaglandins by bovine epithelial and stromal cells stimulated with LPS (Herath et al. 2006). Therefore, we aimed to test whether the stage of the oestrus cycle or exogenous ovarian steroids might affect the innate immune response in the bovine endometrium using ex vivo studies to avoid confounding effects of humoral factors and adaptive immunity in vivo.
This study tested the hypothesis that cyclical variation in ovarian hormone concentrations alters cytokine and chemokine responses in bovine endometrial ex vivo organ cultures (EVOCs) and purified cell populations challenged with LPS or E. coli. Two main approaches were used. First, inflammatory responses to E. coli or LPS were examined in tissues collected from animals at different stages of the oestrus cycle. Secondly, tissues and cells were treated with exogenous oestradiol and progesterone or with inhibitors of the oestradiol or progesterone receptors (PRs). Comparisons were made to the glucocorticoid dexamethasone, which is an established modulator of innate immune responses (Kern et al. 1988, Waage & Bakke 1988).
Materials and methods
Organ and cell culture
Uteri with no gross evidence of genital disease or microbial infection were collected over a 10-month period from postpubertal mixed-breed beef heifers or dairy cows within 15 min of slaughter at a commercial slaughterhouse, as part of the routine operation of the slaughterhouse. Cattle up to 120 days post partum were not used to avoid confounding experiments due to the presence of ubiquitous bacterial contamination and disruption of the epithelium, which is typical of the puerperal endometrium (Herath et al. 2009b, Wathes et al. 2009). The beef heifers (n=174) were 20–26 months old, reared on extensive grassland and had never been pregnant or inseminated. Dairy cows that were pregnant, as determined when the uterine horns were opened (see below), were excluded from the study. The stage of reproductive cycle was determined by examination of ovarian morphology and vasculature, as described previously in detail (Ireland et al. 1979) and by the measurement of hormones in peripheral blood. In accordance with these criteria, stage I is defined as the period from days 1 to 4 of the oestrus cycle, stage II as the period from days 5 to 10, stage III as the period from days 11 to 17 and stage IV as the period from days 18 to 20. Only animals that had gross evidence of ovarian cyclic activity were included. To further evaluate the stage of ovarian cycle, blood samples were collected from the animal carcass at the time of uteri collection, allowed to clot at room temperature and then centrifuged at 2000 g for 15 min to separate the serum, which was then divided in aliquots of 1.5 ml in Eppendorf tubes and frozen at −80 °C until used for progesterone analysis (see below). Within this study, animals from stages IV and I of the oestrus cycle were grouped together, as this represents the follicular phase when serum progesterone concentration is <1 ng/ml.
The uteri were kept on ice for ∼1 h until further processing at the laboratory. Endometrial tissue for EVOC was collected from the contralateral horn, unless otherwise stated, and the intercaruncular areas of the endometrium, except for comparison of responses between intercaruncular and caruncular tissue, using sterile 8 mm-diameter biopsy punches (Stiefel Laboratories Ltd, High Wycome, UK), as described previously (Borges et al. 2012). Tissues were cultured in 24-well plates (TPP, Trasadingen, Switzerland) containing 2 ml complete medium/well, and comprised the following: phenol red-free RPMI 1640 medium (Sigma–Aldrich) containing 10% heat inactivated, double charcoal-stripped, foetal bovine serum (Biosera, East Sussex, UK). The EVOC treatments (see below) were initiated within 4 h of slaughter and maintained in a humidified, 5% CO2 in air atmosphere incubator at 37 °C, with supernatants collected after 6, 24 or 48 h.
Endometrial cells were isolated as described previously (Cronin et al. 2012, Turner et al. 2014). The epithelial and stromal cells were cultured in a complete medium and plated at a density of 1×105 cells/ml in 24-well plates (TPP). The purity of epithelial and stromal cell populations was confirmed by cell morphology and flow cytometric analysis of cytokeratin and vimentin expression respectively (Fortier et al. 1988, Turner et al. 2014).
The isolation and culture of peripheral blood mononuclear cells (PBMCs) were performed as described previously (Herath et al. 2007, Amos et al. 2014). Cells were seeded into 24-well plates at a density of 1×106 cells/well in 1 ml of complete medium and the medium was changed every 2 days until cells exhibited the characteristic macrophage morphology (Steinman & Cohn 1973). The cell population phenotype, which was CD14+, CD45+ and MHC class II+, was confirmed by flow cytometry as described previously (Herath et al. 2007, Price et al. 2013).
Cultures of E. coli (isolate MS499) obtained from an animal with persistent uterine disease, and identified as an endometrial pathogenic E. coli (Sheldon et al. 2010, Goldstone et al. 2014), were grown overnight in Luria-Bertani medium (Sigma–Aldrich). Bacteria were re-suspended to 1×108 colony forming units (CFU)/ml in sterile PBS (Life Technologies Ltd), followed by centrifugation at 6000 g for 10 min at 4 °C. After washing, bacteria were diluted to 1×103 CFU/ml in a complete medium ready for experimental use. Ultrapure LPS from E. coli 0111:B4 was obtained from Invivogen (Toulouse, France). Ovarian and glucocorticoid steroids (E2, E2758; progesterone, P8783 and dexamethasone, D4902) and the steroid receptor antagonists (MPP dihydrochloride hydrate, M7068 and mifepristone, M8046) were obtained from Sigma–Aldrich Ltd. Ovarian and glucocorticoid steroids were prepared by dissolving 1 mg of the steroid in 1 ml absolute ethanol. Stock solutions were prepared at a concentration of 20 μg/ml in complete medium. Final concentrations of E2, progesterone and dexamethasone were prepared by further dilutions in a complete medium. The final concentration of ethanol within tissue or cell cultures was equal to or less than one part in 200 000. MPP dihydrochloride hydrate was prepared by dissolving 20 mg in 1 ml dimethyl sulfoxide. A stock solution was prepared at a concentration of 1 mg/ml (1.85 mM) in complete medium. The final concentration of MPP dihydrochloride hydrate (100 nM) was prepared by further dilution in complete medium. The final concentration of dimethyl sulfoxide within tissue of cell cultures was one part in 370 000. Mifepristone was prepared by dissolving 10 mg in 1 ml absolute ethanol. A stock solution was prepared at a concentration of 1 mg/ml (2.32 mM) in a complete medium. The final concentration of mifepristone (100 nM) was prepared by further dilution in complete medium. The final concentration of ethanol within tissue of cell cultures was one part in 232 000. All treatments were carried out in complete medium that did not contain antibiotics to ensure that bacteria were alive and replicating.
Validation of innate immune responses of EVOCs
To compare endometrial innate immune responses between beef heifers and dairy cows, EVOCs were prepared from beef heifer (n=9) and dairy (n=7) cow uteri that were within the early luteal phase of the oestrus cycle. Comparison of EVOCs using endometrial tissue from intercaruncular (n=4) and caruncular (n=4) zones of the endometrium was performed using early-luteal-phase beef heifer uteri. Endometrial EVOCs were treated with the control medium or a medium containing 1 μg/ml LPS or 1×103 CFU/ml E. coli for 24 h. Comparison of EVOCs using endometrial tissue from the horn ipsilateral (n=13) and contralateral (n=29) to the active corpus luteum (CL) structure was performed using early-luteal-phase beef heifer uteri. Endometrial EVOCs were treated with the control medium or a medium containing 1 μg/ml LPS for 24 h. After treatment, supernatants were collected and stored at −20 °C for analysis of IL1β, IL6 and IL8 by ELISA. The EVOC tissues were weighed and stored in 0.5 ml TRI Reagent at −20 °C until RNA extraction and analysis of IL1B, IL6 and CXCL8 (IL8) mRNA expression by quantitative PCR (qPCR).
Uterine innate immune responses and stage of the oestrus cycle
To evaluate the effect of the stage of the oestrus cycle, tissues from 55 beef heifers were divided into three groups by examination of ovarian morphology (Ireland et al. 1979) and retrospective serum progesterone analysis (see below): follicular phase was defined by ovarian stage (Ireland et al. 1979, stages IV and I) with a serum progesterone concentration of <1 ng/ml (n=6), early luteal phase (Ireland et al. 1979, stage II) with a serum progesterone concentration of 1–2 ng/ml (n=10) and mid-luteal phase (Ireland et al. 1979, stage III) with a serum progesterone concentration of >2 ng/ml (n=39). Endometrial EVOCs from each group were treated with the control medium or a medium containing 1 μg/ml LPS or 1×103 CFU/ml E. coli for 6 or 24 h. After treatment, the supernatants were collected and stored at −20 °C for analysis of IL1β, IL6, IL8 and PGE2 by ELISA. The EVOC tissues were weighed and stored in 0.5 ml TRI Reagent at −20 °C until RNA extraction and analysis of IL1B, IL6, CXCL8 and CCL5 mRNA expression by qPCR.
Steroids and innate immune responses within ex vivo organ cultures
A total of 76 beef heifers were used to evaluate the effect of ovarian and glucocorticoid steroids on innate immune responses within the bovine endometrium. Endometrial EVOCs were divided into two groups according to a retrospective analysis of serum progesterone concentration: group 1 with a serum progesterone concentration of <2 ng/ml (n=20) and group 2 with a serum progesterone concentration of >2 ng/ml (n=56). Endometrial EVOCs were pre-treated for 24 h with the control medium or a medium containing 3 pg/ml oestradiol, 5 ng/ml progesterone or 5 ng/ml dexamethasone. The concentrations of oestradiol and progesterone reflect serum concentrations approximately at the time of ovulation and during the luteal phase of the oestrus cycle respectively (Sheldon et al. 2002, Jimenez-Krassel et al. 2009, Scully et al. 2014). The concentration of dexamethasone used was based upon the recommended potency range from the manufacturer (4–500 ng/ml) and a previous publication, with the aim being to use a minimal effective dose (Kern et al. 1988). After 24 h, EVOCs were challenged with the control medium or a medium containing 1 μg/ml LPS or 1×103 CFU/ml E. coli for a further 24 h in the presence of the steroids. After challenge, supernatants were collected and stored at −20 °C for analysis of IL1β, IL6 and IL8 by ELISA. The EVOC tissues were weighed and stored in 0.5 ml TRI Reagent at −20 °C for analysis of IL1B, IL6, CXCL8 and CCL5. In addition, the expression of ESR1 and PGR mRNA was determined by qPCR to ensure that the tissue remained responsive to exogenous steroids for the 48 h treatment period.
Steroids and cellular innate immune responses
To evaluate the impact of ovarian and glucocorticoid steroids on in vitro cellular innate immune responses, endometrial epithelial and stromal cells and PBMCs, collected from four beef heifers, were pre-treated for 24 h with the control medium or a medium containing 3 pg/ml oestradiol, 5 ng/ml progesterone or 5 ng/ml dexamethasone. After 24 h, cells were challenged with the control medium or a medium containing 100 ng/ml LPS for a further 24 h in the presence of the steroids. After challenge, supernatants were collected and stored at −20 °C for analyses of IL6 and IL8 by ELISA. Additionally, endometrial stromal cells from three beef heifers were pre-treated for 24 h with the control medium or a medium containing 1–30 pg/ml oestradiol, 1–30 ng/ml progesterone or 5 ng/ml dexamethasone. After 24 h, cells were challenged with the control medium or a medium containing 100 ng/ml LPS for a further 24 h in the presence of the steroids. After challenge, supernatants were collected and stored at −20 °C for analysis of IL6 and IL8 by ELISA. Cell survival was assessed by the mitochondria-dependent reduction of 3-(4,5-dimethyltiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) to formazan, as described previously (Mosmann 1983). Briefly, supernatants were removed and replaced with fresh complete medium containing 0.5 mg/ml MTT before being incubated with cells at 37 °C in a humidified atmosphere of 5% CO2 for 1 h. The medium was then removed and cells were washed with sterile PBS before lysis with dimethyl sulfoxide and measurement of the optical density was performed at a wavelength of 570 nm using a microplate reader (POLARstar Omega; BMG Labtech, Offenburg, Germany). The correlation between MTT OD 570 measurements and the number of live cells was confirmed using Trypan blue exclusion and counting the number of live cells using a haemocytometer.
Steroid receptor antagonists and endometrial innate immune responses
To further explore the impact of steroids on immunity in the endometrium, their actions were inhibited using antagonists for their nuclear receptors. Endometrial EVOCs from 23 beef heifers in the luteal phase of the oestrus cycle were pre-treated for 24 h with the control medium or a medium containing 5 ng/ml progesterone, 3 pg/ml oestradiol or 5 ng/ml dexamethasone. Pre-treatments were performed in the presence or absence of the oestrogen receptor alpha (ERα) antagonist MPP dihydrochloride hydrate (MPP, 100 nM) (Sun et al. 2002) or the PR/glucocorticoid receptor (GR) antagonist mifepristone (100 nM) (Skinner et al. 1999, Siemieniuch et al. 2010). After 24 h, the EVOCs were challenged with the control medium or a medium containing 1 μg/ml LPS or 1×103 CFU/ml E. coli for a further 24 h in the presence of the steroids and/or antagonists. After challenge, the supernatants were collected and stored at −20 °C for analysis of IL1β, IL6 and IL8 by ELISA, and EVOC tissues were weighed.
Enzyme immune assays
Concentrations of IL1β, IL6 and IL8 in EVOC and cell culture supernatants were measured in duplicate by ELISA according to the manufacturer's instructions (Bovine IL1β Screening Set (ESS0027; Thermo Fisher Scientific, Perbio Science UK Ltd, Cramlington, UK); Bovine IL6 Screening Set (ESS0029; Thermo Fisher Scientific) and Human CXCL8/IL8 DuoSet (DY208; R&D Systems Europe Ltd, Abingdon, UK)). The human CXCL8/IL8 DuoSet has previously been validated for the measurement of bovine IL8 (Rinaldi et al. 2008). To take into account the differences between the weights of EVOC tissues, concentrations are expressed in picograms per milligram of tissue. Serum progesterone concentrations were determined by a Progesterone Enzyme Immunoassay (Ridgeway Research Ltd, St Briavels, UK), according to the manufacturer's instructions. The inter- and intra-assay coefficient of variation values were all <12 and 7% respectively; the limits of detection were 12.5 pg/ml for IL1β, 75.0 pg/ml for IL6, 5.7 pg/ml for IL8 and 0.1 ng/ml for progesterone.
Gene expression analysis
Gene expression analysis was performed according to the MIQE guidelines (Bustin et al. 2009). Total RNA was isolated from EVOC tissues by homogenising the tissue in 2 ml tubes containing 0.5 ml TRI Reagent (Sigma–Aldrich) and lysing matrix D (MP Biomedicals, Cambridge, UK) at a rate of 6.0 m/s for 2 min. After homogenisation, tubes were centrifuged at 12 000 g for 10 min, the supernatants were transferred to fresh 2 ml Eppendorf tubes, and RNA extraction from TRI Reagent was then performed according to the manufacturer's instructions. RT of 1 μg mRNA was performed in a 20 μl reaction volume using the QuantiTect RT Kit (Qiagen), according to the manufacturer's instructions.
QPCR for IL1B, CXCL8, CCL5, ESR1, GAPDH and ACTB was performed by multiplex probe-based PCR and comprised two panels of primer/probe combinations (panel 1: IL1B, CXCL8 and GAPDH and panel 2: CCL5, ESR1 and ACTB), which simultaneously measured cDNA for each target gene. PCR primers and probes were designed using the Eurofins MWG Operon qPCR primer/probe design software (https://ecom.mwgdna.com/services/webgist/dual_probe_design?usca_p=t) and validated by BLAST analysis against the Bos taurus (taxid:9913) Refseq mRNA database. Primers/probes were obtained from Eurofins MWG Operon (Ebersberg, Germany). Multiplex qPCR was performed in 10 μl reaction volumes comprising 1× QuantiFast Multiplex PCR Master Mix (Qiagen) with primers and probes added in nuclease-free water to a final concentration of 0.4 and 0.2 μM, respectively, and 2 μl of cDNA. Thermal cycling parameters were as follows: one cycle of 95 °C for 5 min followed by 50 cycles of 95 °C for 15 s and 60 °C for 30 s.
QPCR for IL6 and PGR was performed by SYBR Green-based PCR because existing primers/methods for these genes were already present in the laboratory. PCR primers were designed using the Eurofins MWG Operon qPCR primer/probe design software (https://ecom.mwgdna.com/services/webgist/dual_probe_design?usca_p=t) and validated by BLAST analysis against the Bos taurus (taxid:9913) Refseq mRNA database. PGR and IL6 primers were obtained from Eurofins MWG Operon and Sigma–Aldrich respectively. SYBR Green-based PCR was performed in a 25 μl reaction volume comprising 1× QuantiFast SYBR Green PCR Master Mix (Qiagen) with primers added in nuclease-free water to a final concentration of 0.4 μM and 2 μl of cDNA. Thermal cycling parameters were as follows: one cycle of 95 °C for 5 min, followed by 40 cycles of 95 °C for 30 s and 60 °C for 60 s.
All primers and probes used are given in detail in Table 1. The expression of each gene was normalised against the geometric mean of the reference genes GAPDH and ACTB, which were invariant across treatment groups (Vandesompele et al. 2002), and the relative quantification method was employed to quantify target gene mRNA within samples (Nolan et al. 2006). To generate standard curves, total RNA extracted from EVOC tissues that had been treated with 1 μg/ml LPS for 24 h was reverse transcribed to cDNA, as described. Tenfold serial dilutions of this reference cDNA were prepared (neat to 1×10−5) in nuclease-free water (Qiagen). For each sample, target and reference gene mRNA abundance was determined from the appropriate standard curve (quantification cycle (Cq)). Changes in mRNA abundance between samples were then determined from the ratio of the target gene Cq to the reference gene Cq.
Quantitative PCR primers and probes used for gene expression analysis.
Statistical analyses were performed using SAS version 8.0 with the animal as the experimental unit. Initially, the data were tested for homogeneity, followed by analysis using the general linear model (GLM) multiplex ANOVA using Dunnett's pairwise multiple comparison t-test for individual group comparisons. Gene data are expressed as dot plots, protein data are expressed as histograms and data are expressed as mean with s.e.m. P<0.05 was considered statistically significant.
Endometrial EVOCs from beef heifers and dairy cows respond similarly to LPS and E. coli
To validate the utility of the EVOC system, selected cytokine and chemokine responses were compared between EVOCs from beef heifer and dairy cow endometria. Endometrial EVOCs from beef heifers and dairy cows accumulated more IL1β, IL6 and IL8 following challenge with LPS or E. coli compared with the control medium (Fig. 1A, B and C, P<0.0001). There was also increased IL1B, IL6 and CXCL8 mRNA expression in response to challenge with LPS or E. coli (Fig. 1D, E and F, P<0.0001). However, there was no significant difference in the protein or mRNA responses to LPS or E. coli between EVOCs from beef heifers and dairy cows.
To further validate the use of EVOCs, inflammatory responses to E. coli and LPS were compared between caruncular and intercaruncular EVOCs collected from beef heifer endometrium. Endometrial EVOCs from both areas secreted more IL1β, IL6 and IL8 in response to challenge with LPS or E. coli compared with the control medium (Fig. 1G, H and I, P<0.05). However, more IL1β, IL6 and IL8 were secreted from intercaruncular EVOCs compared with caruncular EVOCs following challenge with LPS (P<0.05), and more IL1β and IL6 were secreted following challenge with E. coli (P<0.05).
Finally, inflammatory responses to LPS were compared between endometrial EVOCs collected from the horn ipsilateral and contralateral to the active CL structure. Endometrial EVOCs from both horns secreted more IL6 in response to challenge with LPS compared with the control medium (Fig. 1K, P<0.001), and there was a trend for increased IL1β and IL8. Importantly, however, there were no significant differences in responses to LPS between EVOCs from the ipsilateral or contralateral horns.
Endometrial innate immune responses and stage of the oestrus cycle
To investigate the role of the oestrus cycle in modulating innate immunity, responses to challenge with LPS or E. coli for 6 or 24 h were tested using EVOCs of intercaruncular endometrium collected from beef animals at different stages of the oestrus cycle. As expected, cows in the mid-luteal phase had more progesterone present in their serum (8.51+0.75 ng/ml, P<0.001) compared with cows in the follicular (0.08+0.07 ng/ml) or early luteal (0.5+0.21 ng/ml) phases.
Following a 6 h challenge with LPS or E. coli, EVOCs accumulated more IL6 (P<0.0001), compared with the control medium, and after a 24 h challenge with LPS or E. coli, EVOCs accumulated more IL1β, IL6 and IL8 (P<0.0001) (Fig. 2). Mid-luteal-phase cow EVOCs accumulated more IL6 (P<0.05), but not IL1β or IL8, in response to a 6 h challenge with LPS compared with follicular- and early-luteal-phase cow EVOCs (Fig. 2A, B and C). There were no significant differences between the stages of the oestrus cycle following a 6 h challenge with E. coli. Mid-luteal-phase cow EVOCs accumulated less IL1β (P<0.05) compared with early-luteal-phase cow EVOCs following a 24 h challenge with E. coli, but there was no difference in IL6 or IL8 secretion (Fig. 2D, E and F). There were no significant differences between stages of the oestrus cycle for EVOCs challenged with LPS for 24 h. To further explore the impact of the stage of the oestrus cycle on inflammatory responses in the endometrium, the EVOC data were combined (follicular, n=19; early-luteal, n=43 and mid-luteal, n=22). Treatment with LPS or E. coli increased the accumulation of IL6 compared with the control (305.7±13.1, 243.4±12.4 and 93.1±13.9 pg/ml, respectively; ANOVA, P<0.001). However, there was no significant effect of the stage of cycle (P=0.23) or the interaction between treatment and stage (P=0.88).
Challenge of EVOCs with E. coli or LPS for 6 or 24 h also increased the expression of IL1B, IL6, CXCL8 and CCL5 (Fig. 3, P<0.05). However, there were no consistent differences in mRNA expression among the different stages of the oestrus cycle measured for any of the four genes examined at 6 or 24 h (Fig. 3).
Ovarian steroids and endometrial responses to LPS or E. coli
In the absence of an effect of the stage of the oestrus cycle on the inflammatory response, an alternative approach was examined to test whether ovarian steroids regulate endometrial innate immunity, by treating EVOCs with exogenous steroids before challenge with LPS or E. coli. The EVOCs were retrospectively divided into two groups, based on the serum progesterone concentration: <2 ng/ml (follicular) and >2 ng/ml (luteal). As ovarian hormones regulate the ESR1 and PGR genes (Kimmins & MacLaren 2001), to confirm the responsiveness of the EVOCs to steroid treatment, the impact of oestradiol, progesterone and the glucocorticoid dexamethasone on ESR1 and PGR mRNA expression was measured. Within EVOCs of follicular-phase endometrial tissue, 48 h treatment with E2 increased the expression of ESR1 (Fig. 4A, P<0.05) and PGR (Fig. 4C, P<0.05) mRNA compared with the control, and progesterone significantly decreased the expression of PGR (Fig. 4C, P<0.05) mRNA, while dexamethasone had no effect (Fig. 4A and C). The same pattern of change in expression of PGR mRNA was measured in EVOCs obtained from luteal-phase cows (Fig. 4D, P<0.05), but none of the steroids had any significant effect on ESR1.
As mentioned previously, irrespective of steroid treatment, EVOCs responded to challenge with LPS by accumulating IL6 and IL8, and accumulated IL6, IL8 and IL1β in response to challenge with E. coli (Fig. 5, P<0.05). However, pre-treatment of EVOCs with either oestradiol or progesterone for 24 h had no significant effect on responses to subsequent challenges with LPS or E. coli, and there was no effect of the prior hormone concentration in the animals. However, dexamethasone reduced the accumulation of IL1β in EVOCs before and following challenge with LPS or E. coli (Fig. 5A and B, P<0.05). As mentioned previously, challenge of EVOCs with LPS or E. coli increased the expression of IL1B, IL6, CXCL8 and CCL5 mRNA compared with the control medium (Fig. 6, P<0.001). However, pre-treatment with oestradiol or progesterone had no effect on mRNA abundance in response to these challenges, and again this was irrespective of the prior hormone concentration in the animals. By contrast, pre-treatment with dexamethasone reduced the expression of IL1B (Fig. 6A and B, P<0.05) and CXCL8 (Fig. 6 E and F, P<0.05) in EVOCs before and following challenge with LPS or E. coli, and CCL5 (Fig. 6G and H, P<0.05) following challenge with LPS or E. coli, compared with the control medium.
To determine whether the lack of effect of ovarian steroids on endometrial innate immune responses was unique to the EVOCs, experiments were also conducted using pure populations of endometrial cells. Endometrial epithelial cells, stromal cells and PBMCs accumulated IL6 (Fig. 7A, B and C) and IL8 (Fig. 7D, E and F) in response to challenge with LPS for 24 h (P<0.001). However, pre-treatment with oestradiol or progesterone for 24 h had no significant effect on IL6 or IL8 production of each of the three cell types. However, pre-treatment with dexamethasone reduced the accumulation of IL6 in epithelial cells (Fig. 7A, P<0.05) and IL8 in stromal cells and PBMCs (Fig. 7E and F, P<0.05).
However, different concentrations of exogenous steroid could modulate the cellular responses and, hence, endometrial stromal cells were treated with a range of concentrations of oestradiol or progesterone, before challenge with LPS. Stromal cells were used for two reasons: first, stromal cells are more responsive to LPS than epithelial cells, and the increased dynamic range gave the best chance of observing an effect. Secondly, epithelial cells are lost during and after parturition exposing the underlying stromal cells to bacterial infection (Archbald et al. 1972). Pre-treatment of endometrial stromal cells with 1–30 pg/ml oestradiol (Fig. 8A and C) or 1–30 ng/ml progesterone (Fig. 8B and D) for 24 h, did not modulate IL6 or IL8 responses to LPS during subsequent challenges.
Ovarian steroid receptor antagonists and endometrial responses to LPS or E. coli
To examine whether steroid nuclear receptor function modulates endometrial innate immunity, EVOCs were pre-treated with the ERα antagonist MMP or the PR/GR antagonist mifepristone, with or without the appropriate steroid present. After the 24 h pre-treatment, EVOCs were challenged with the control medium or a medium containing LPS or E. coli. Endometrial EVOCs accumulated IL1β, IL6 and IL8 in response to challenge with LPS and IL1β and IL6 following challenge with E. coli (Fig. 9, P<0.0001). Pre-treatment of EVOCs for 24 h with oestradiol and/or MPP had no effect on endometrial responses to challenge with LPS or E. coli (Fig. 9A, D and G). Pre-treatment for 24 h with progesterone and/or mifepristone also had no significant effect on endometrial responses to challenge with LPS or E. coli (Fig. 9B, E and H). However, pre-treatment with dexamethasone reduced (P<0.05) the accumulation of IL1β in response to challenge with E. coli, and importantly, pre-treatment with dexamethasone and mifepristone blocked the IL1β-inhibiting effect of dexamethasone (Fig. 9C, P<0.05).
In vivo, there is clear evidence for a protective effect of oestradiol or oestrus against infection of the uterus, and for a disease-promoting effect of progesterone or the luteal phase of the oestrus cycle (Rowson et al. 1953, Del Vecchio et al. 1992, Lewis 2004). Although, several explanations for these effects have been explored previously, the mechanistic explanations are elusive, particularly in relation to leukocyte population differences and neutrophil function (Subandrio & Noakes 1997, Subandrio et al. 2000, Winters et al. 2003). Thus, we reasoned that ovarian steroids might modulate innate immune responses in the endometrium. However, in this study, the stage of the oestrus cycle did not influence the cytokine or chemokine response of ex vivo endometrial tissue to E. coli or LPS at the gene or the protein level. Furthermore, exogenous ovarian steroids did not modulate the innate immune response by endometrial tissue or cells. Finally, even blocking the nuclear receptors for oestradiol or progesterone did not affect the inflammatory response to E. coli or LPS. We conclude that ovarian steroids have little effect on in vitro inflammatory responses associated with innate immunity in the bovine endometrium.
The use of EVOCs maintains the architecture of cells in the tissue and retains an imprint of the stage of the oestrus cycle of the animal. Using EVOCs also avoids potential confounders of in vivo studies, including humoral factors, effects of nutrition and adaptive immune responses, enabling exploration of the impact of steroids in the localised tissue and cells of the endometrium, independent of the whole-animal response. In this study, endometrial EVOCs from beef heifers were a good surrogate for tissues from dairy cows, producing similar cytokine and chemokine responses to E. coli and LPS. Furthermore, using tissue and cells from beef heifers, potential confounders in dairy cows, such as insemination, pregnancy, previous uterine disease and lactation, were removed. Yet, the increased cytokine and chemokine secretion and increased mRNA expression following challenge of EVOCs with E. coli or LPS mirror the changes during disease in vivo (Herath et al. 2009b, Sheldon et al. 2009). Additionally, EVOCs collected from the horn ipsilateral or contralateral to the active corpus luteum were equally responsive to LPS, suggesting that the inflammatory response is not modulated by differing concentration gradients of the hormone across the two horns. This view is supported by gene array analyses, which report hundreds of differentially expressed genes in the endometrium of luteal- vs follicular-phase animals, but very few genes differ in expression between the horn ipsilateral and contralateral to the corpus luteum, and those that do have very low ratios (Bauersachs et al. 2005, Shimizu et al. 2010). However, EVOCs incorporating tissue from intercaruncular areas of the endometrium were more responsive to challenge with LPS or E. coli than caruncular tissue. With over 1100 differentially expressed genes between intercaruncular and caruncular tissues, including several inflammation and immune-regulating genes (Mansouri-Attia et al. 2009), use of tissues from the intercaruncular zones was an important optimisation step.
Central to the response to bacterial challenge is the detection of pathogen-associated molecular patterns by TLRs, and in particular for E. coli infection, binding of LPS by TLR4. Endometrial epithelial and stromal cells also express TLRs, including TLR4, and produce IL6 and IL8 following challenge with LPS (Herath et al. 2006, Sheldon & Roberts 2010). In this study, E. coli and LPS stimulated the accumulation of IL1β and IL8 by 24 h. The kinetics of IL6 production probably reflects the roles of IL6 in the early response to infection such as leukocyte recruitment, B-lymphocyte development, antibody secretion by plasma cells and the regulation of acute-phase proteins. Interleukin 8 (IL8), a potent chemo-attractor and activator of neutrophils and T-lymphocytes, is secreted by monocytes, lymphocytes, fibroblasts and epithelial and endothelial cells (Mukaida 2000, Schaefer et al. 2004). IL1β is secreted predominantly by monocytes following inflammasome activation and stimulates the production of additional pro-inflammatory cytokines, such as IL6 (van de Veerdonk et al. 2011), and chemokines such as IL8, which recruit more immune cells and promote phagocytosis and bacterial clearance (Petrilli et al. 2007).
The most striking observations in this study were that endometrial tissue and cell responses to challenge with E. coli or LPS were not influenced by the stage of the oestrus cycle, or by the addition of exogenous oestradiol or progesterone. First, we found that the stage of the oestrus cycle did not affect innate immune responses of endometrial EVOCs challenged with LPS or E. coli. Hence, we considered whether separating out the cellular populations would uncover steroid-responsive effects, using differential regulation of ESR1 and PGR mRNA to verify that the endometrial cells were responsive to exogenous oestradiol and progesterone. However, ovarian steroids had no effect on separated endometrial cell or PBMC responses to challenge with LPS or E. coli, and similarly, inhibition of ERα or PR had no effect on innate immune responses; although, inhibiting the GR inhibited dexamethasone-related inflammatory modulation. One could argue that the initial staging of the oestrus cycle may have been erroneous. However, more than 150 animals were used across the studies, peripheral plasma progesterone concentrations were used to verify the stage of cycle, and the variance across groups was small irrespective of the stage of cycle. It could also be argued that higher steroid concentrations in the uterine tissue, compared with the peripheral plasma, might effectively modulate inflammatory responses (McCracken et al. 1984, Weems et al. 1988, Einer-Jensen et al. 1989). However, extended dose range experiments showed no effect on inflammatory responses to LPS, and EVOC tissue had been exposed to native uterine steroid concentrations. Taken together, these data suggest that there is neither a direct effect of the ovarian steroids on innate immunity nor oestrus cyclic regulation of ovarian steroid receptor expression likely to affect innate immunity.
In vivo, there is a clear oestrus-dependent effect on basal mRNA expression of cytokines and chemokines such as IL1B, CXCL8 and CXCL5 in cells collected from the uterine lumen (Fischer et al. 2010). Thus, this gives rise to a question as to how the negative results of this study in vitro fit into the whole animal effects? First, there may be an innate immune effect mediated by regulatory molecules not investigated in this study. In vitro, exogenous ovarian steroids reduce the synthesis of prostaglandin F2α and prostaglandin E2 in endometrial cells (Herath et al. 2006). Other classes of molecules, such as antimicrobial peptides, lipoxins or resolvins, could also be examined. Secondly, alteration of the adaptive immune response would have a significant effect on disease outcome. The presence of ovarian steroid receptors on immune cells suggests the possibility of their regulation, and there is evidence from humans that ovarian steroids have a significant impact on disease outcome (Waage et al. 1990, Wira et al. 2005, Rodriguez-Garcia et al. 2013a,b). Thirdly, there may be an indirect effect of ovarian steroids on innate or adaptive immunity. Indeed, in this study, dexamethasone reduced IL1β, IL6 and IL8 secretion, together with IL1B, IL6, CXCL8 and CCL5 mRNA responses to challenge with E. coli or LPS. In addition, the GR antagonist mifepristone attenuated the suppressive effect of dexamethasone on IL1β secretion following E. coli challenge. Kuse et al. have recently demonstrated a regulatory effect of ovarian steroids on NR3C1 expression within the bovine endometrium. As a function of the stage of the oestrus cycle, NR3C1 expression within the bovine endometrium was greater during the mid-luteal phase when progesterone concentrations are high, compared with other phases of the oestrus cycle, and the glucocorticoid cortisol more strongly suppressed PGF2α production during the mid-luteal phase than during the follicular phase. The addition of progesterone to cultured endometrial epithelial cells also increased expression of NR3C1, while oestradiol reduced expression levels (Kuse et al. 2013). A future approach might also probe single cell responses to PAMPs and steroids as, although this study has shown no response to ovarian steroids within large populations of cells in vitro, individual cells may respond. Indeed, such an approach has recently revealed the production of the lymphosteroid pregnenolone by Th2 T cells, which is associated with immunosuppression, inhibiting Th cell proliferation and B cell immunoglobulin class switching (Mahata et al. 2014).
In conclusion, there was no effect of the stage of the oestrus cycle, exogenous ovarian steroids or inhibition of their nuclear receptors on key cytokine and chemokine responses to E. coli or LPS in endometrial tissues or cells. The lack of effect of ovarian steroids challenges the central dogma that steroids suppress immunity across species.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
This work was supported by funding from the UK Biotechnology and Biological Sciences Research Council (BB/1017240/1), FAPEMIG and CNPq fellow (Conselho Nacional de Desenvolvimento Científico e Tecnológico – grant numbers PDE 201916/2012-6 and PDE 200885/2010-3).
The authors thank James Cronin for isolation and culture of PBMCs, and James Cronin and Steve Jeremiah for their assistance with ELISA.
AmosMRHealeyGDGoldstoneRJMahanSMDüvelASchuberthHJSandraOZiegerPDieuzy-LabayeISmithDG2014Differential endometrial cell sensitivity to a cholesterol-dependent cytolysin links Trueperella pyogenes to uterine disease in cattle. Biology of Reproduction9054. (doi:10.1095/biolreprod.113.115972)
BauersachsSUlbrichSEGrossKSchmidtSEMeyerHHEinspanierRWenigerkindHVermehrenMBlumHSinowatzF2005Gene expression profiling of bovine endometrium during the oestrous cycle: detection of molecular pathways involved in functional changes. Journal of Molecular Endocrinology34889–908. (doi:10.1677/jme.1.01799)
CroninJGTurnerMLGoetzeLBryantCESheldonIM2012Toll-like receptor 4 and MYD88-dependent signaling mechanisms of the innate immune system are essential for the response to lipopolysaccharide by epithelial and stromal cells of the bovine endometrium. Biology of Reproduction8651–59. (doi:10.1095/biolreprod.111.092718)
FischerCDrillichMOdauSHeuwieserWEinspanierRGablerC2010Selected pro-inflammatory factor transcripts in bovine endometrial epithelial cells are regulated during the oestrous cycle and elevated in case of subclinical or clinical endometritis. 22818–829. (doi:10.1071/RD09120)
HerathSLillySTSantosNRGilbertROGoetzeLBryantCEWhiteJOCroninJSheldonIM2009bExpression of genes associated with immunity in the endometrium of cattle with disparate postpartum uterine disease and fertility. Reproductive Biology and Endocrinology755. (doi:10.1186/1477-7827-7-55)
Jimenez-KrasselFFolgerJKIrelandJLSmithGWHouXDavisJSLonerganPEvansACIrelandJJ2009Evidence that high variation in ovarian reserves of healthy young adults has a negative impact on the corpus luteum and endometrium during estrous cycles in cattle. Biology of Reproduction801272–1281. (doi:10.1095/biolreprod.108.075093)
MahataBZhangXKolodziejczykAAProserpioVHaim-VilmovskyLTaylorAEHebenstreitDDinglerFAMoignardVGottgensB2014Single-cell RNA sequencing reveals T helper cells synthesizing steroids de novo to contribute to immune homeostasis. Cell Reports71130–1142. (doi:10.1016/j.celrep.2014.04.011)
PriceJCBromfieldJJSheldonIM2013Pathogen-associated molecular patterns initiate inflammation and perturb the endocrine function of bovine granulosa cells from ovarian dominant follicles via TLR2 and TLR4 pathways. Endocrinology1543377–3386. (doi:10.1210/en.2013-1102)
SheldonIMRycroftANDoganBCravenMBromfieldJJChandlerARobertsMHPriceSBGilbertROSimpsonKW2010Specific strains of Escherichia coli are pathogenic for the endometrium of cattle and cause pelvic inflammatory disease in cattle and mice. PLoS ONE5e9192. (doi:10.1371/journal.pone.0009192)
TurnerMLCroninJGHealeyGDSheldonIM2014Epithelial and stromal cells of bovine endometrium have roles in innate immunity and initiate inflammatory responses to bacterial lipopeptides in vitro via Toll-like receptors TLR2, TLR1 and TLR6. Endocrinology1551453–1465. (doi:10.1210/en.2013-1822)